Hardware Accelerated Python Based Finite Element Analysis of Reinforced Concrete Member

This paper presents a nonlinear finite element analysis (FEA) of RC beams by adopting material models based on machine learning (ML).The Gaussian Process Regression (GPR) approach is considered for constructing concrete and steel material models.However, a GPR material model has an increased computational load, so it is difficult to use in the nonlinear analysis of RC structures composed of numerous members.To solve this limitation, GPU acceleration is based on the constitution of the parallelized computing structure.GPU-accelerated Python-based FEA program is developed to trace the nonlinear behaviour of RC beams.The FEA and experimental data for two representative RC beams are compared.The results obtained from the developed program confirm that the solution procedure using GPU acceleration with GPR material models can effectively be used in the nonlinear analysis of large RC structures with nonlinear behaviours.